Functional neuroimaging is an extremely promising and rapidly developing discipline that involves the use of advanced technologies for non-invasive imaging of the active human brain. Current techniques attempt to correlate patterns of neural activation with the performance of cognitive tasks that are designed to isolate specific brain functions. The purpose of our project is to develop new methods for the analysis of functional magnetic resonance imaging (fMRI) data. Our research currently focuses on three aspects of the analysis: image registration in the Fourier domain, statistical inference under spatio-temporal models of the activation process, and advanced parallel and vector processing, including adaptive, real-time analysis of fMRI data. Our methods will also supply to other imaging modalities.